wouldn't that http://plugcomputer.org/plugwiki/index.php/Enabling_UBIFS answer your question?

Why I'm lost: if I search for "netlink fd: Protocol not supported" I only get 5 matches (of which one is this thread .

Two of them (e.g. https://forums.openfiler.com/viewtopic.php?id=5210) recommend to use an older kernel - strange!? Maybe I'll give at try ...

thx pingtoo, but as in http://plugcomputer.org/plugforum/index.php?topic=2329.msg13231#msg13231 it tells me:

root@Share:~# ietd
netlink fd: Protocol not supported

But can't figure out what that means ...

All,

though I gave up for the time being, I'd like to understand what you mean with this:


Do you mean cbxbiker61's kernels are not suitable for the Ubuntu 9.04 version which comes with the plug? I've updated both the kernels in flash and on SD-card with his versions, where latter is the normal distro I'm using (booting from flash is only fallback). I'm considering to give the Debian Squeeze a chance and did some trial, however it will still take me some time to set it up properly and to migrate my existing setup [1].


? udev is installed already ?

[1] luckily I have a second spare plug, which I bought to have a backup to be save if the first dies. I want to work in parallel on this one, but when done I just want to swap SD-cards (and amend/modify uboot-settings). I recall that the system "learns" it's MAC address and saves it somewhere, which needs to be deleted before moving cards, but I can't recall which file it is...

Thx sfzhi, I just did the test on my SheevaPlug with the following (odd) result:

My U-Boot version is: U-Boot 1.1.4 (Dec 27 2009 - 22:03:21) Marvell version: 3.4.27 - pingtoo patch.01

During my first attempt I "deleted" the arguement for the ethaddr environment variable, saved the settings and booted the system. Hence it came up with the right address. My first guess was: "Hurray, it's working", but found this to be the wrong conclusion later, because:

When I did a second reboot, the system came up with no eth interface at all, just the loopback!?

So I went into U-Boot again to see what happened and was surprised to find (printenv) the ethaddr env-var set as

ethaddr=00:50:43:a6:0c:23

So I started the system over and again, but again no eth int.

Back in U-Boot the ethaddr arg was different again and seems to be changing randomly from reboot/reset to reset.

ethaddr=00:50:43:98:04:26
ethaddr=00:50:43:b6:36:29
ethaddr=00:50:43:79:23:2f

So my guess what happened: during my first attempt I just altered the U-Boot settings after the system had already been initialized and as I just issued a boot rather than reset command it came up with the expected MAC addr.

However something is still odd with the U-Boot:
i) though the ethaddr is null in the saved config(?) the u-boot printenv command delivers a random result.
ii) however the u-boot is not using that random MAC address to initialize the interface and thus the OS shows no eth int!?

Only after I set the ethaddr again the plug comes up with that interface enabled.
This can be very misleading as printenv gives you an address though none is set.

Now finding this it reminded me of that strange behavior I run in before, but which gets a _bit_ clearer, now. (http://plugcomputer.org/plugforum/index.php?topic=1304.msg8181#msg8181)

PS: Update in short how it "works"

Marvell>> setenv ethaddr
Marvell>> printenv ethaddr
## Error: "ethaddr" not defined
Marvell>> saveenv
Saving Environment to NAND...
Erasing Nand...Writing to Nand... done
Marvell>> reset
[...]
Marvell>> printenv ethaddr
ethaddr=00:50:43:42:34:37
Marvell>> printenv ethaddr
ethaddr=00:50:43:42:34:37
Marvell>> reset
[...]
Marvell>> printenv ethaddr
ethaddr=00:50:43:ea:2d:38

and so on ...

PS2: I only have an eth0 if I set the "right" MAC address in u-boot.

Just being curious: is it really required to set the MAC address in U-Boot!? (I also set the MAC to the info on my plug's sticker after flashing to a new U-Boot release, but currently hard for me to test as my plug is productive sitting in the basement far away from a console :-()
Of course setting a wrong/foreign MAC might cause issues, however what MAC is used if no one is set (00:...)?
Looking at other systems they normally use their native MAC if nothing is defined/overruled?!
Of course if you flash U-Boot it comes with a predefined default MAC address setting, but what will happen if you remove it
if allowed at all?!

hmm, in his reply http://plugcomputer.org/plugforum/index.php?topic=2333.msg13249#msg13249 cbxbiker61 said about crc32c: "That's built-in."

Though I seam to have iscsi_trgt now

root@Share:/lib/modules/2.6.35.7# lsmod
Module                  Size  Used by
dm_crypt               11267  0
dm_mod                 55673  1 dm_crypt
ipv6                  233263  26
hmac                    2411  0
sha1_generic            1673  0
mv_cesa                 9178  0
root@Share:/lib/modules/2.6.35.7# modprobe -l | grep isc 
kernel/drivers/misc/eeprom/eeprom_93cx6.ko
kernel/drivers/scsi/scsi_transport_iscsi.ko
kernel/drivers/scsi/libiscsi.ko
kernel/drivers/scsi/libiscsi_tcp.ko
kernel/drivers/scsi/iscsi_tcp.ko
kernel/drivers/scsi/iscsitarget/iscsi_trgt.ko
kernel/drivers/usb/misc/emi26.ko
kernel/drivers/usb/misc/emi62.ko
kernel/drivers/usb/misc/sisusbvga/sisusbvga.ko
kernel/sound/usb/misc/snd-ua101.ko

I can't find a hint about crc32c

root@Share:/lib/modules/2.6.35.7# modprobe -l | grep crc
kernel/lib/crc-itu-t.ko

Thus because it's build in?

Another FYI, after installing the new kernel I once again ran "sh -x /etc/init.d/iscsitarget start".

However when I spot for any changes the output is still the same as in http://plugcomputer.org/plugforum/index.php?topic=2329.msg13222#msg13222

root@Share:~# diff s1 s2 ! s1 is a copy of the above post's output while s2 is copy of an up2date "sh -x /etc/init.d/iscsitarget start" amended with the "tracking string" you see below
96a97
> !Is this only difference?

lost!?

Linux Share 2.6.35.7 #4 PREEMPT Fri Oct 1 21:46:30 MDT 2010 armv5tel

Another q: would it be good enough to have the kernel's system.map file in the /boot dir?
re https://secure.wikimedia.org/wikipedia/en/wiki/System.map SVGALib expects to find it /lib/modules/$(uname -r)/build/System.map?

However on my system it's a link to a non-existing directory!?

root@Share:/boot# ls -al /lib/modules/2.6.35.7/build
lrwxrwxrwx 1 root root 51 Oct  2 07:55 /lib/modules/2.6.35.7/build -> /home/kelly/src/Sheeva/Kernel/linux-2.6.35.7-sheeva

As in http://plugcomputer.org/plugforum/index.php?topic=2333.msg13243#msg13243 cbxbiker61 added the iscsi_tgt support to 2.6.35.7, though it's there now the start fails still.
I followed up and asked for crc32c support please, too.

thx a lot, I think one step further, but still failing to start.
However as birdman spottet in the other thread would it also need crc32c enabled?

sorry, for my noob q, however re my iSCSI issue http://plugcomputer.org/plugforum/index.php?topic=2329.0 would you know if your kernel supports iSCSI target?
--
thx & cheers
Oliver

sure, however reading your other advises my (non existent) U*X knowledge told me that those additional ones would only work if at least the ietd would start (which it doesn't)?!?

how can I check this?

thx :-)

PS:

root@Share:~# lsmod
Module                  Size  Used by
dm_crypt               11267  0
dm_mod                 55673  1 dm_crypt
ipv6                  233263  26
hmac                    2411  0
sha1_generic            1673  0
mv_cesa                 9178  0
root@Share:~# modprobe -l | grep isc
kernel/drivers/misc/eeprom/eeprom_93cx6.ko
kernel/drivers/usb/misc/emi26.ko
kernel/drivers/usb/misc/emi62.ko
kernel/drivers/usb/misc/sisusbvga/sisusbvga.ko
kernel/sound/usb/misc/snd-ua101.ko

thx pingtoo

Ah, I guess one step further thou I don't know, yet what this wants to tell me ;-)

root@Share:~# /usr/libexec/ietd --uid=root --gid=root --port=3260
-bash: /usr/libexec/ietd: No such file or directory
root@Share:~# which ietd
/usr/sbin/ietd
root@Share:~# /usr/sbin/ietd  --uid=root --gid=root --port=3260
netlink fd: Protocol not supported

hmm, are you're using cbxbiker61's kernels? BTW: I'm currently using his  2.6.35.7 #2 PREEMPT Wed Sep 29 16:46:15 MDT 2010 version

Ah, sorry, think got your intention, now. Here's the full "dump". As I'm in a hurry I just had a brief look at it (of course there are a couple if checks e.g. n != y or equivalent, but I couldn't spot anything right away)


root@Share:~# sh -x /etc/init.d/iscsitarget start
+ PID_FILE=/var/run/iscsi_trgt.pid
+ CONFIG_FILE=/etc/ietd.conf
+ DAEMON=/usr/sbin/ietd
+ PATH=/sbin:/bin:/usr/sbin:/usr/bin
+ MEM_SIZE=1048576
+ . /lib/lsb/init-functions
+ FANCYTTY=
+ [ -e /etc/lsb-base-logging.sh ]
+ . /etc/lsb-base-logging.sh
+ ISCSITARGET_ENABLE=false
+ ISCSITARGET_DEFAULTS_FILE=/etc/default/iscsitarget
+ [ -s /etc/default/iscsitarget ]
+ . /etc/default/iscsitarget
+ ISCSITARGET_ENABLE=true
+ RETVAL=0
+ [ true = true ]
+ ietd_start
+ log_daemon_msg Starting iSCSI enterprise target service
+ [ -z Starting iSCSI enterprise target service ]
+ log_use_usplash
+ [ n = y ]
+ type usplash_write
+ log_to_console log_daemon_msg Starting iSCSI enterprise target service
+ [ n != y ]
+ [ no != yes ]
+ readlink /proc/self/fd/0
+ stdin=/dev/pts/0
+ [ /dev/pts/0 != /dev/pts/0 ]
+ return 0
+ log_use_fancy_output
+ TPUT=/usr/bin/tput
+ EXPR=/usr/bin/expr
+ [ -t 1 ]
+ [ xxterm !=  ]
+ [ xxterm != xdumb ]
+ [ -x /usr/bin/tput ]
+ [ -x /usr/bin/expr ]
+ /usr/bin/tput hpa 60
+ /usr/bin/tput setaf 1
+ [ -z ]
+ FANCYTTY=1
+ true
+ /usr/bin/tput xenl
+ /usr/bin/tput cols
+ COLS=80
+ [ 80 ]
+ [ 80 -gt 6 ]
+ /usr/bin/expr 80 - 7
+ COL=73
+ printf  * Starting iSCSI enterprise target service
 * Starting iSCSI enterprise target service       + /usr/bin/expr 80 - 1
+ /usr/bin/tput hpa 79
                                                                               + printf
 + configure_memsize
+ [ -e /proc/sys/net/core/wmem_max ]
+ echo 1048576
+ [ -e /proc/sys/net/core/rmem_max ]
+ echo 1048576
+ [ -e /proc/sys/net/core/wmem_default ]
+ echo 1048576
+ [ -e /proc/sys/net/core/rmem_default ]
+ echo 1048576
+ [ -e /proc/sys/net/ipv4/tcp_mem ]
+ echo 1048576 1048576 1048576
+ [ -e /proc/sys/net/ipv4/tcp_rmem ]
+ echo 1048576 1048576 1048576
+ [ -e /proc/sys/net/ipv4/tcp_wmem ]
+ echo 1048576 1048576 1048576
+ modprobe -q crc32c
+ RETVAL=1
+ [ 1 != 0 ]
+ log_end_msg 1
+ [ -z 1 ]
+ log_use_usplash
+ [ n = y ]
+ type usplash_write
+ log_to_console log_end_msg 1
+ [ n != y ]
+ [ no != yes ]
+ readlink /proc/self/fd/0
+ stdin=/dev/pts/0
+ [ /dev/pts/0 != /dev/pts/0 ]
+ return 0
+ [ 73 ]
+ [ -x /usr/bin/tput ]
+ printf \r
+ /usr/bin/tput hpa 73
                                                                         + [ 1 -eq 0 ]
+ printf [
[+ /usr/bin/tput setaf 1
+ printf fail
fail+ /usr/bin/tput op
+ echo ]
]
+ return 1
+ exit 1

@Pingtoo: I tried it with my own domain first, but as that didn't work as well I tried to stick as closely to an example as possible.

@Birdman: think I have no related messages in syslog

The output of "sh -x /etc/init.d/iscsitarget start" shows something like
...
[+ /usr/bin/tput setaf 1
+ printf fail
fail+ /usr/bin/tput op
+ echo ]
...

Re running the script step-by-step will me take some time - I have to understand it's structure first ;-)